ACL RD-TEC 1.0 Summarization of A00-2040
Paper Title:
A FINITE STATE AND DATA-ORIENTED METHOD FOR GRAPHEME TO PHONEME CONVERSION
A FINITE STATE AND DATA-ORIENTED METHOD FOR GRAPHEME TO PHONEME CONVERSION
Primarily assigned technology terms:
- 5-fold cross-validation
- algorithm
- approximation
- cross-validation
- data-oriented method
- database
- exhaustive search
- finite state
- finite state automata
- finite state transducer
- grapheme-to-phoneme conversion
- hyphenation
- indexing
- learning
- longest-match replacement
- morphology
- parsing
- part-of-speech tagging
- pc-kimmo
- phonetic transcription
- processing
- prolog
- regular expression
- rule ordering
- rule sampling
- sampling
- scoring
- search
- segmentation
- shallow parsing
- speech synthesis
- spelling
- spelling correction
- spelling correction system
- synthesis
- tagging
- text to speech
- text-to-speech
- text-to-speech system
- tile
- tokenization
- transcription
- transducer
- transducers
- transformation-based learning
Other assigned terms:
- 10-fold cross-validation
- abbreviation
- abbreviations
- alignment problem
- alignment procedure
- alphabet
- approach
- automata
- boundary marker
- case
- characters
- composition
- conversion rule
- cpu time
- data sets
- dictionary
- disjunction
- dutch
- edit distance
- error rate
- experimental results
- fact
- fsa utilities
- grapheme
- implementation
- input string
- interpretation
- knowledge
- lexical database
- linguistic
- linguistic knowledge
- mapping
- mappings
- maps
- method
- methodology
- morphological rules
- part-of-speech
- phoneme
- phoneme sequence
- phoneme string
- phonemes
- probabilities
- probability
- procedure
- process
- prolog implementation
- pronunciation
- regular expressions
- rule set
- segments
- statistics
- substring
- syllables
- symbols
- syntax
- tags
- technology
- term
- test set
- text
- tile context
- training
- training data
- training material
- training set
- transcriptions
- transformation
- transformation rules
- translations
- word
- words
